Mark Suster Is Wrong, You Should Be A Startup Entrepreneur(collaborable.com)
collaborable.com
Mark Suster Is Wrong, You Should Be A Startup Entrepreneur
http://collaborable.com/blog/mark-suster-is-wrong-you-should-be-a-startup-entrepreneur
6 comments
I agree that Mark's article is full of great advice, and that he is not explicitly answering the question for anyone.
This blog was my genuine reaction to Mark's list of characteristics, which is something that in the past I may have read and felt discouraged.
It's a shame to discourage when I now know the truth: that none of those characteristics must be identified in order to get started and eventually succeed at a startup.
This blog was my genuine reaction to Mark's list of characteristics, which is something that in the past I may have read and felt discouraged.
It's a shame to discourage when I now know the truth: that none of those characteristics must be identified in order to get started and eventually succeed at a startup.
I'm amazed you took the leap with zero savings and a two-month old baby to support. Congrats; what is this 'financially rewarding' startup you created?
My entrepreneurial story isn't over, but I started around 3 years ago with an ecommerce niche store (I'll get into specifics on my blog soon).
I make twice what I did before I took the leap, and that's because most of our resources go back into the business.
I make twice what I did before I took the leap, and that's because most of our resources go back into the business.
I like the sentiment. I too think that anyone who has an entrepreneurial spark should give it a go (with proper preparation of course.) It also seems folly that if someone isn't predisposed to certain attributes, they shouldn't attempt the launch.
However, if you're going to throw down a gauntlet with a linkbait title like that, don't wuss out like this: "Okay, maybe I should say “I disagree” with Mark on this issue"
However, if you're going to throw down a gauntlet with a linkbait title like that, don't wuss out like this: "Okay, maybe I should say “I disagree” with Mark on this issue"
Haha, thanks for the feedback, maybe next time I'll up the hard-assery :)
I encourage it. I have a theory that all good meritorious technology communities are strengthened by mild aggression towards members.
By which I mean things like "Why would you EVER think this was a good idea?" not "You're fat, ugly and I fucking HATE your skank of a wife".
By which I mean things like "Why would you EVER think this was a good idea?" not "You're fat, ugly and I fucking HATE your skank of a wife".
I would love to hear feedback on this blog from entrepreneurs and wantrepreneurs alike.
http://xkcd.com/386/
'nuff said.
Suster's remarks are similar to many others for starting a business in information technology, software, Web 2,0, social media, etc.
For all these remarks there is a very simple response: In the US, coast to coast, villages, towns, cities, on Main Street there are successful businesses in everything from pizza carryout, Chinese food carryout, franchised fast food, portrait photography, jewelry, fashion boutiques, dentistry, medical family practice, CPA, wines and spirits, auto parts, auto repair, auto body repair, gasoline, convenience store, plumber, electrician, plumbing supply, electrical supply, hardware, building supply, on and on. A significant fraction of these Main Street business are big truck -- little truck, that is, buy in large quantities at a low price per unit and sell in small quantities at a high price per unit.
These businesses are a huge fraction of the US economy. The fraction is so large that there is no chance that any significant fraction of the business founders could satisfy any list of personality characteristics as specific as Suster's.
Yes, nearly none of these Main Street businesses will become worth over $1 billion (McDonald's, Domino's, maybe a few more did). Still, the businesses are successful, e.g., commonly last decades and for an owner pay to support a family.
So, what about the businesses Suster is considering in Web 2.0, etc.?
Well, these businesses have some big advantages including Moore's law and the Internet and commonly can be started for much less money than even a carryout pizza shop or, really, nearly any of the Main Street businesses.
E.g., the people who mow grass in my neighborhood drive nice, new trucks, each pulling a low volume, handmade trailer carrying a high end lawn mower worth something over $10,000. Altogether they arrive to mow some grass with capital expense ruining $40,000 which now would buy one heck of a first-cut Web 2.0 server farm.
So, setting aside Suster's focus on businesses that can be started for the cost of a computer and a good Internet connection and a some amount of hacker time and soon be worth over $500 million, just why should the advantages of Moore's law and the Internet require such particular and rare personality characteristics when millions of Main Street businesses without these advantages do not?
The answer is simple: The business based on computing don't require satisfying any such list of particular and rare personality characteristics.
So what's going on? Suster is working way too hard looking for the secrets and, thus, is getting way off the track.
For all these remarks there is a very simple response: In the US, coast to coast, villages, towns, cities, on Main Street there are successful businesses in everything from pizza carryout, Chinese food carryout, franchised fast food, portrait photography, jewelry, fashion boutiques, dentistry, medical family practice, CPA, wines and spirits, auto parts, auto repair, auto body repair, gasoline, convenience store, plumber, electrician, plumbing supply, electrical supply, hardware, building supply, on and on. A significant fraction of these Main Street business are big truck -- little truck, that is, buy in large quantities at a low price per unit and sell in small quantities at a high price per unit.
These businesses are a huge fraction of the US economy. The fraction is so large that there is no chance that any significant fraction of the business founders could satisfy any list of personality characteristics as specific as Suster's.
Yes, nearly none of these Main Street businesses will become worth over $1 billion (McDonald's, Domino's, maybe a few more did). Still, the businesses are successful, e.g., commonly last decades and for an owner pay to support a family.
So, what about the businesses Suster is considering in Web 2.0, etc.?
Well, these businesses have some big advantages including Moore's law and the Internet and commonly can be started for much less money than even a carryout pizza shop or, really, nearly any of the Main Street businesses.
E.g., the people who mow grass in my neighborhood drive nice, new trucks, each pulling a low volume, handmade trailer carrying a high end lawn mower worth something over $10,000. Altogether they arrive to mow some grass with capital expense ruining $40,000 which now would buy one heck of a first-cut Web 2.0 server farm.
So, setting aside Suster's focus on businesses that can be started for the cost of a computer and a good Internet connection and a some amount of hacker time and soon be worth over $500 million, just why should the advantages of Moore's law and the Internet require such particular and rare personality characteristics when millions of Main Street businesses without these advantages do not?
The answer is simple: The business based on computing don't require satisfying any such list of particular and rare personality characteristics.
So what's going on? Suster is working way too hard looking for the secrets and, thus, is getting way off the track.
just why should the advantages of Moore's law and the Internet require such particular and rare personality characteristics when millions of Main Street businesses without these advantages do not?
Well, mostly because starting an internet business is a lot harder, because there's (more or less) a single global market and it's more competitive.
If I have the only sporting goods store in Iron Knob it doesn't really matter that there's a much bigger and better one in Medicine Hat. But if I start a crappy online bookstore it's going to have a lot of trouble competing with Amazon.
It's a winner-takes-all market, so in order to succeed you need to be doing something that nobody else is doing, or else doing it better than everyone else, or at least doing it the same as everyone else but having significantly better traction. That means that most internet startups are aimed towards doing stuff that nobody has ever done before, which requires a lot more hard work and brainpower than ordering tennis balls by the thousand and selling them by the can.
The upside of the winner-takes-all marketplace is that the winner takes all. The downside of the winner-takes-all marketplace is that the winner takes all.
Now I suppose there are some unoriginal business opportunities on the internet, but they're largely a bit sleazy -- buying up domains and sticking SEO-optimized crap on them, for instance. Even that is a lot harder than buying a lawnmower and knocking on the doors of folks with messy lawns though.
Well, mostly because starting an internet business is a lot harder, because there's (more or less) a single global market and it's more competitive.
If I have the only sporting goods store in Iron Knob it doesn't really matter that there's a much bigger and better one in Medicine Hat. But if I start a crappy online bookstore it's going to have a lot of trouble competing with Amazon.
It's a winner-takes-all market, so in order to succeed you need to be doing something that nobody else is doing, or else doing it better than everyone else, or at least doing it the same as everyone else but having significantly better traction. That means that most internet startups are aimed towards doing stuff that nobody has ever done before, which requires a lot more hard work and brainpower than ordering tennis balls by the thousand and selling them by the can.
The upside of the winner-takes-all marketplace is that the winner takes all. The downside of the winner-takes-all marketplace is that the winner takes all.
Now I suppose there are some unoriginal business opportunities on the internet, but they're largely a bit sleazy -- buying up domains and sticking SEO-optimized crap on them, for instance. Even that is a lot harder than buying a lawnmower and knocking on the doors of folks with messy lawns though.
You gave a good explanation of a major point I omitted, the common situation of Main Street businesses having a strong geographical barrier to entry.
And your point about "winner take all" on the Internet is simple but basically correct.
Okay, but still: Joe has an idea, maybe Facebook for cat owners, optimized for mobile with a special alarm when GPS says they are near a cat owner of the opposite sex, a pivot from his semi-successful MySpace for dog owners, a pivot from his quasi-successful MeetUp for gerbil lovers. To Joe, each of these is a really new, unique, no competition idea. And "romantic angle?" "Nearly all romantic angle." And, if he gets just 2% of all the ...!
Maybe Joe's ideas are terrific, and maybe they suck. If he does fail fast, then for each such pivot he will soon learn which case is true.
So, all that said, just where in there does Joe need all those special personality characteristics Suster listed? I mean, Joe has a brother Sam who started a Chicago style pizza place in LA -- that was his innovative idea for a Main Street business. Like nearly all Main Street business owners, Sam is just a regular guy, just like his brother Joe.
Net, I see Moore's law and the Internet as mostly just advantages except for the one issue you mentioned, the flip side of the great opportunity of serving the world is no geographical barrier to entry. So, if Joe's idea really is new and good and he is on the way to tens of millions of uniques a month just from his cable TV ISP connection and some mid-tower case servers on wire shelf racks in his spare bedroom, then I'm missing just why Joe has to have personality characteristics different from the rest of Main Street. Or, if Joe's latest pivot idea basically sucks, then the solution is to think of a better idea and not have some rare personality characteristics.
Uh, Suster is right in the soup of Internet social media entrepreneurship taking in ideas, thoughts, impressions, etc. from a fire hose and pushing them out at the same rate. Or, Fred Wilson once commented that he doesn't see how Suster gets any sleep. But my view of Suster is that, however successful he has been or is being, he's not the most disciplined thinker. Of course, with the fire hose rates he is using, even if only some of his ideas are good then he can do well, which I suspect he is.
In a sense, I cheated on Suster: I turned his context around to essentially just another Main Street business except exploiting Moore's law and the Internet (and as you pointed out, also needs some real novelty to make up for the lack of geographical barrier to entry) from his context of, say, Mark Zuckerberg as he crosses the stormy seas of growth to 600 million users, Steve Jobs when he executed his semi-triumphant role of return of the native, Andrew Mason growing Groupon and telling Google that they didn't have money enough, etc.
Both contexts are relevant: Of course, Suster wants $1 billion or so exits, but a lot of entrepreneurs would be happy owning over 50% of a $50 million exit.
Maybe Suster is saying that for a $1+ billion exit, or $50+ billion, need a Facebook, and for that have to go through some wild and crazy times, and for that need a Zuckerberg, and for that need Suster's rare personality characteristics.
I still don't buy it.
Here's another approach: A calm, very carefully reasoned guy gets told in high school that his goal of going to MIT is a hoot. But he goes. Basically he gets done in three years but sticks around for a fourth year and, then, takes two reading courses. One of these is in Lie Groups and Lie Algebras from I. Singer. Then he goes to grad math at Berkeley and cleans up a really messy proof for a major result in differential geometry. He is back at MIT for a while, then goes to military work at the Institute for Defense Analysis at Princeton, gets fired for giving Newsweek an interview about the Viet Nam war, then becomes Chairman of the math department at Stony Brook. There he works with S. Chern and gets a new result that now physicists like for their work with differential geometry.
Then some earlier little efforts in business pay off, and he invests the money and plays the stock market. He does well, in two years or so multiplies his money by a factor of 12. He leaves to start a hedge fund and does better and better. Finally he takes human judgment out of the trading and does it all with software based on math ideas and carefully back tested. Soon he is paying himself about $2 billion a year. Last year he retired and now is doing philanthropy and is back to math.
Uh, Brookhaven National Lab wanted to collide some heavy ions, see if they could create a quark-gluon plasma, needed $12 million or so; Congress wouldn't fund it; so he did.
Lesson: He's very successful and is not a wild or crazy guy.
Sure, he's Jim Simons, COB of Renaissance Technologies on Long Island as in his biographical lecture at MIT
http://paul.kedrosky.com/archives/2011/01/james_simons_sp.ht...
and his testimony before Waxman's committee with article at
http://www.nytimes.com/2008/11/14/business/14hedge.html
and Simons's video at
http://video.nytimes.com/video/2008/11/13/business/119483281...
I. Singer? He's as in the Atiyah-Singer index theorem which won one of the first, maybe the first, Abel Prizes created by the Nobel committee as a Nobel prize in math.
Uh, for success on the Internet, we agree on the need for powerful, new ideas, hopefully with a barrier to entry from difficult to duplicate or equal technology and not just from devoted users, and for that we need to be quite open about where the ideas come from. Uh, Simons is one of a kind in hedge funds. Well Internet entrepreneurship needs to be welcoming one of a kind people, also, and Suster needs to be looking for them; then Suster needs to throw out his list of personality characteristics.
Drawing from the first Indiana Jones movie, with his list Suster is "digging in the wrong place". His filter is just independent of anything that makes any sense and, thus, is throwing both wheat and chaff indiscriminately.
Beyond that, there's another push on Suster: Recently John Doerr spouted off again on the old saw that "Ideas are easy, and execution is everything.". No: Bad ideas are easy, and then execution is everything. Good ideas are difficult, and then execution is routine.
Well, for evaluating ideas as new, correct, significant, and powerful with valuable results for business, can count without taking shoes off all the information technology venture partners who can do such evaluations: Uh just doing well evaluating new, correct, and significant takes a peer-reviewed journal editor, and there may not be even ten information technology venture partners in all of the US qualified to be a journal editor.
So, basically the venture partners ignore the ideas and, instead, say, for a seed round look at the user interface and evaluate if people might like it, for a Series A look at ComScore numbers, for a Series B look at revenue and earnings. Basically their view is that at each stage, doing well on their criteria makes also looking at the idea unnecessary. But this does not mean that ideas are irrelevant or that good ideas are easy. Still, from this old saw, Suster goes too far and expects that execution will be some mad house where he wants his personality characteristics. If he wants a mad house, then he can have it.
And your point about "winner take all" on the Internet is simple but basically correct.
Okay, but still: Joe has an idea, maybe Facebook for cat owners, optimized for mobile with a special alarm when GPS says they are near a cat owner of the opposite sex, a pivot from his semi-successful MySpace for dog owners, a pivot from his quasi-successful MeetUp for gerbil lovers. To Joe, each of these is a really new, unique, no competition idea. And "romantic angle?" "Nearly all romantic angle." And, if he gets just 2% of all the ...!
Maybe Joe's ideas are terrific, and maybe they suck. If he does fail fast, then for each such pivot he will soon learn which case is true.
So, all that said, just where in there does Joe need all those special personality characteristics Suster listed? I mean, Joe has a brother Sam who started a Chicago style pizza place in LA -- that was his innovative idea for a Main Street business. Like nearly all Main Street business owners, Sam is just a regular guy, just like his brother Joe.
Net, I see Moore's law and the Internet as mostly just advantages except for the one issue you mentioned, the flip side of the great opportunity of serving the world is no geographical barrier to entry. So, if Joe's idea really is new and good and he is on the way to tens of millions of uniques a month just from his cable TV ISP connection and some mid-tower case servers on wire shelf racks in his spare bedroom, then I'm missing just why Joe has to have personality characteristics different from the rest of Main Street. Or, if Joe's latest pivot idea basically sucks, then the solution is to think of a better idea and not have some rare personality characteristics.
Uh, Suster is right in the soup of Internet social media entrepreneurship taking in ideas, thoughts, impressions, etc. from a fire hose and pushing them out at the same rate. Or, Fred Wilson once commented that he doesn't see how Suster gets any sleep. But my view of Suster is that, however successful he has been or is being, he's not the most disciplined thinker. Of course, with the fire hose rates he is using, even if only some of his ideas are good then he can do well, which I suspect he is.
In a sense, I cheated on Suster: I turned his context around to essentially just another Main Street business except exploiting Moore's law and the Internet (and as you pointed out, also needs some real novelty to make up for the lack of geographical barrier to entry) from his context of, say, Mark Zuckerberg as he crosses the stormy seas of growth to 600 million users, Steve Jobs when he executed his semi-triumphant role of return of the native, Andrew Mason growing Groupon and telling Google that they didn't have money enough, etc.
Both contexts are relevant: Of course, Suster wants $1 billion or so exits, but a lot of entrepreneurs would be happy owning over 50% of a $50 million exit.
Maybe Suster is saying that for a $1+ billion exit, or $50+ billion, need a Facebook, and for that have to go through some wild and crazy times, and for that need a Zuckerberg, and for that need Suster's rare personality characteristics.
I still don't buy it.
Here's another approach: A calm, very carefully reasoned guy gets told in high school that his goal of going to MIT is a hoot. But he goes. Basically he gets done in three years but sticks around for a fourth year and, then, takes two reading courses. One of these is in Lie Groups and Lie Algebras from I. Singer. Then he goes to grad math at Berkeley and cleans up a really messy proof for a major result in differential geometry. He is back at MIT for a while, then goes to military work at the Institute for Defense Analysis at Princeton, gets fired for giving Newsweek an interview about the Viet Nam war, then becomes Chairman of the math department at Stony Brook. There he works with S. Chern and gets a new result that now physicists like for their work with differential geometry.
Then some earlier little efforts in business pay off, and he invests the money and plays the stock market. He does well, in two years or so multiplies his money by a factor of 12. He leaves to start a hedge fund and does better and better. Finally he takes human judgment out of the trading and does it all with software based on math ideas and carefully back tested. Soon he is paying himself about $2 billion a year. Last year he retired and now is doing philanthropy and is back to math.
Uh, Brookhaven National Lab wanted to collide some heavy ions, see if they could create a quark-gluon plasma, needed $12 million or so; Congress wouldn't fund it; so he did.
Lesson: He's very successful and is not a wild or crazy guy.
Sure, he's Jim Simons, COB of Renaissance Technologies on Long Island as in his biographical lecture at MIT
http://paul.kedrosky.com/archives/2011/01/james_simons_sp.ht...
and his testimony before Waxman's committee with article at
http://www.nytimes.com/2008/11/14/business/14hedge.html
and Simons's video at
http://video.nytimes.com/video/2008/11/13/business/119483281...
I. Singer? He's as in the Atiyah-Singer index theorem which won one of the first, maybe the first, Abel Prizes created by the Nobel committee as a Nobel prize in math.
Uh, for success on the Internet, we agree on the need for powerful, new ideas, hopefully with a barrier to entry from difficult to duplicate or equal technology and not just from devoted users, and for that we need to be quite open about where the ideas come from. Uh, Simons is one of a kind in hedge funds. Well Internet entrepreneurship needs to be welcoming one of a kind people, also, and Suster needs to be looking for them; then Suster needs to throw out his list of personality characteristics.
Drawing from the first Indiana Jones movie, with his list Suster is "digging in the wrong place". His filter is just independent of anything that makes any sense and, thus, is throwing both wheat and chaff indiscriminately.
Beyond that, there's another push on Suster: Recently John Doerr spouted off again on the old saw that "Ideas are easy, and execution is everything.". No: Bad ideas are easy, and then execution is everything. Good ideas are difficult, and then execution is routine.
Well, for evaluating ideas as new, correct, significant, and powerful with valuable results for business, can count without taking shoes off all the information technology venture partners who can do such evaluations: Uh just doing well evaluating new, correct, and significant takes a peer-reviewed journal editor, and there may not be even ten information technology venture partners in all of the US qualified to be a journal editor.
So, basically the venture partners ignore the ideas and, instead, say, for a seed round look at the user interface and evaluate if people might like it, for a Series A look at ComScore numbers, for a Series B look at revenue and earnings. Basically their view is that at each stage, doing well on their criteria makes also looking at the idea unnecessary. But this does not mean that ideas are irrelevant or that good ideas are easy. Still, from this old saw, Suster goes too far and expects that execution will be some mad house where he wants his personality characteristics. If he wants a mad house, then he can have it.
Because many of those markets aren't really markets.
In all likelihood, the market of Facebook for cat owners is no different than the market for Facebook. Why would a cat-owning user prefer a dedicated site over Facebook itself? They lose their social network, they have to check multiple websites to stay up to date, and the website owner doesn't have the benefits of scale that FaceBoook does. There's no competitive advantage to that particular idea.
The characteristics that Suster list are important because this is the norm in winner-take-all markets. The vast majority of things that you think are viable markets aren't really; if they were, a competitor would already have gotten to them. So you have to strike out and do something new over and over again, until you finally hit on something that actually works. That's really stressful, particularly when you're doing it without knowing where your next paycheck is coming from.
In all likelihood, the market of Facebook for cat owners is no different than the market for Facebook. Why would a cat-owning user prefer a dedicated site over Facebook itself? They lose their social network, they have to check multiple websites to stay up to date, and the website owner doesn't have the benefits of scale that FaceBoook does. There's no competitive advantage to that particular idea.
The characteristics that Suster list are important because this is the norm in winner-take-all markets. The vast majority of things that you think are viable markets aren't really; if they were, a competitor would already have gotten to them. So you have to strike out and do something new over and over again, until you finally hit on something that actually works. That's really stressful, particularly when you're doing it without knowing where your next paycheck is coming from.
Yes, Facebook for cat owners is absurd; it was supposed to be a hoot and a parody of a silly idea.
Your description of trying over and over is common but should not be the only way.
Broadly another way is really the old entrepreneurship paradigm: Find a nasty problem that many people have and that has no good solution and where a large fraction of the people would be eager to have a good solution.
Look for a good solution. If actually find one, then go forward, implement the solution, and offer it to the people who wanted it.
Then there's the old saw that "Whatever you are working on, at least 20 other people are working on the same thing with the same ideas.". Well, this claim is silly: Anyone who has done much peer-reviewed original research can see clearly that, a large fraction of such research, especially once the work has been reviewed and accepted for publication, is unique with no one else doing anything very close.
E.g., once I was working in a research project applying artificial intelligence (AI) to system and network monitoring and management. Well, one of the first needs is doing well detecting problems in real time. Such detection is for either (A) old problems seen before or (B) new problems, zero day, never seen before. Assume that whenever we see a B problem we implement corrections and, thus, convert it to an A problem and solve it so that we never see it again. So, we are left with detecting B problems.
I thought that the AI techniques we were using were junk.
Indeed, clearly, as we monitor, there are two ways to be wrong, (1) a false alarm where we say that the system is sick when it is healthy and (2) a missed detection of a real problem where we say that the system is healthy when it is sick. So, clearly we are now necessarily close to statistical hypothesis testing with Type I error (false alarms) and Type II error (missed detections).
Then we are necessarily close to the classic Neyman-Pearson result on the way, for each rate of false alarms, to get the lowest possible rate of missed detections.
Well the whole field of zero day monitoring had not yet gotten even this far, which is really just a junior level course in mathematical statistics.
So, we want to do a hypothesis test. Okay but for the large literature of such tests, we have two issues: First, from server farms and networks, we can collect data on many variables, not just one. So, we want to be multi-dimensional. Second, especially being multi-dimensional, we have no hope of knowing the probability distribution of, say, a healthy system. So, we want our work to be distribution-free.
Well, can look through the literature, especially, say, E. Lehmann, and find nothing on multidimensional, distribution-free tests.
So, one Saturday I put my feet up and created a large, new family of such tests. I wrote out theorems and proofs to justify what I was doing. I wrote some corresponding software. Then I had some data from a complicated server farm, washed it through my software, and saw that I was getting what my theorems said. Then I did a long series of Monte Carlo tests with some very complicated data; my detection techniques worked just as intended.
So, get to select false alarm rate in advance and get that rate exactly. There is not enough data to get all of Neyman-Pearson, but in a powerful sense, asymptotically, for whatever false alarm rate is selected the techniques give the lowest possible rate of missed detections.
So, my work is progress in zero day monitoring of complex systems and networks. I published the work.
Got to tell you, history since I did that work and published it shows clearly that I was the only person in the world doing anything like what I did.
Is there a business in this, say, to be sold to HP, Microsoft, IBM, EMC, Cisco, or some such? Maybe, but my current project is easier to do and more valuable.
The lesson is, broadly, if really have something new and advanced, the chances that someone else is doing the same thing are small.
Of course, what I'm really talking about is applied math as the crucial, core 'secret sauce' to get a much better solution to a nasty problem and not just routine software for just some intuitive idea.
Then this is a broad area of opportunities: In column A list nasty problems people would like to have solved. In column B list some applied math techniques, old or new, that take in data and spit out results. Then find a good pair, a problem from column A and a solution from column B where the solution is much better than anything else for the problem and likely difficult to duplicate or equal. Now write the corresponding software and proceed with little risk of anyone else doing the same thing. The key is making the project one in applied math, not just computing or computer science.
Where'd I get this paradigm? Sure: The US DoD has been doing such things with great success all the way back to WWII, and I started my career around DC in DoD work.
Can this paradigm work in Web 2.0 and consumer-facing Internet? I do believe so!
Your description of trying over and over is common but should not be the only way.
Broadly another way is really the old entrepreneurship paradigm: Find a nasty problem that many people have and that has no good solution and where a large fraction of the people would be eager to have a good solution.
Look for a good solution. If actually find one, then go forward, implement the solution, and offer it to the people who wanted it.
Then there's the old saw that "Whatever you are working on, at least 20 other people are working on the same thing with the same ideas.". Well, this claim is silly: Anyone who has done much peer-reviewed original research can see clearly that, a large fraction of such research, especially once the work has been reviewed and accepted for publication, is unique with no one else doing anything very close.
E.g., once I was working in a research project applying artificial intelligence (AI) to system and network monitoring and management. Well, one of the first needs is doing well detecting problems in real time. Such detection is for either (A) old problems seen before or (B) new problems, zero day, never seen before. Assume that whenever we see a B problem we implement corrections and, thus, convert it to an A problem and solve it so that we never see it again. So, we are left with detecting B problems.
I thought that the AI techniques we were using were junk.
Indeed, clearly, as we monitor, there are two ways to be wrong, (1) a false alarm where we say that the system is sick when it is healthy and (2) a missed detection of a real problem where we say that the system is healthy when it is sick. So, clearly we are now necessarily close to statistical hypothesis testing with Type I error (false alarms) and Type II error (missed detections).
Then we are necessarily close to the classic Neyman-Pearson result on the way, for each rate of false alarms, to get the lowest possible rate of missed detections.
Well the whole field of zero day monitoring had not yet gotten even this far, which is really just a junior level course in mathematical statistics.
So, we want to do a hypothesis test. Okay but for the large literature of such tests, we have two issues: First, from server farms and networks, we can collect data on many variables, not just one. So, we want to be multi-dimensional. Second, especially being multi-dimensional, we have no hope of knowing the probability distribution of, say, a healthy system. So, we want our work to be distribution-free.
Well, can look through the literature, especially, say, E. Lehmann, and find nothing on multidimensional, distribution-free tests.
So, one Saturday I put my feet up and created a large, new family of such tests. I wrote out theorems and proofs to justify what I was doing. I wrote some corresponding software. Then I had some data from a complicated server farm, washed it through my software, and saw that I was getting what my theorems said. Then I did a long series of Monte Carlo tests with some very complicated data; my detection techniques worked just as intended.
So, get to select false alarm rate in advance and get that rate exactly. There is not enough data to get all of Neyman-Pearson, but in a powerful sense, asymptotically, for whatever false alarm rate is selected the techniques give the lowest possible rate of missed detections.
So, my work is progress in zero day monitoring of complex systems and networks. I published the work.
Got to tell you, history since I did that work and published it shows clearly that I was the only person in the world doing anything like what I did.
Is there a business in this, say, to be sold to HP, Microsoft, IBM, EMC, Cisco, or some such? Maybe, but my current project is easier to do and more valuable.
The lesson is, broadly, if really have something new and advanced, the chances that someone else is doing the same thing are small.
Of course, what I'm really talking about is applied math as the crucial, core 'secret sauce' to get a much better solution to a nasty problem and not just routine software for just some intuitive idea.
Then this is a broad area of opportunities: In column A list nasty problems people would like to have solved. In column B list some applied math techniques, old or new, that take in data and spit out results. Then find a good pair, a problem from column A and a solution from column B where the solution is much better than anything else for the problem and likely difficult to duplicate or equal. Now write the corresponding software and proceed with little risk of anyone else doing the same thing. The key is making the project one in applied math, not just computing or computer science.
Where'd I get this paradigm? Sure: The US DoD has been doing such things with great success all the way back to WWII, and I started my career around DC in DoD work.
Can this paradigm work in Web 2.0 and consumer-facing Internet? I do believe so!
You make some great points here, especially with your last sentence: "So what's going on? Suster is working way too hard looking for the secrets and, thus, is getting way off the track."
A vision and the will to embark on the journey is common to all businesses.
Sometimes the vision is more about the benefits of the business, and in that case, grand success may be a bit more elusive until the entrepreneur learns to focus on the business of business itself.
But again, it would be ridiculous to expect every entrepreneur to be "prepared" before starting a startup -- of any kind.
A vision and the will to embark on the journey is common to all businesses.
Sometimes the vision is more about the benefits of the business, and in that case, grand success may be a bit more elusive until the entrepreneur learns to focus on the business of business itself.
But again, it would be ridiculous to expect every entrepreneur to be "prepared" before starting a startup -- of any kind.
Read his advice, he tells it like it is. If you choose to do it, then you have answered the question for yourself. If Mark advises against it and you think he's right, indeed you're not an entrepreneur. If you do it anyway, simply prove that it's the right answer.